SSDT with large, retail databases?

Question

I'm just getting my feet wet with SSDT because I'm excited about its source control and deployment capabilities. I'm working on an instance of Microsoft Dynamics GP which consists of a couple thousand tables, about 20k stored procs, etc. I won't be modifying the schema very much, instead only working to modify a handful of the procs (probably a couple hundred max).

However, creating a new SSDT project off of our live instance to facilitate this work creates a project that won't build, and the number of errors is fairly massive. Suffice to say correcting 21570 errors won't be feasible for the scope of this project given the relative size of the work we actually mean to do--especially since we didn't author it ourselves to begin with.

I don't want to give up hope, though--I'm curious if it's possible just to use SSDT to house & maintain the procs we're using, and deploy only the scripts we modify even though there's much more to the databases that we don't touch. Is this at all realistic?

Answer 1

Try creating a dacpac of the existing database (use SQLPackage to generate it), then reference that as a DB Reference in your project as "same database". I blogged about using SQLPackage and database references here: http://schottsql.blogspot.com/2012/10/ssdt-external-database-references.html

That way your project will contain just the objects you are creating but still reference the existing objects.